I have read about and used plaid sync API using its cursors. I am curious how plaid internally design its cursor efficiently? I suspect plaid must persist the cursor, but how does it prevent excessive cursor situation. for example: one call sync using an item id with count = 10, plaid creates a cursor for fetching the next page, another call using the same item id with count = 50, plaid has to create another cursor for that call. with enough variation of sync calls, plaid may have to create a cursor for each transaction and it has to handle cursor ordering issue as well. I am curious how it's handled by plaid.
I tried plaid sync APIs with different count value.